INDEX

Explanations

patterns and relationships in logical arrangements involving placements or classifications

oai_token-act-pair · gpt-4o-mini Triggered by @bot

like "middle", "top", "central", and "right"

np_acts-logits-general · gemini-2.0-flash

upper and middle positions

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 20-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_20/width_16k/average_l0_71

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.20.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 فريبيس

-0.84

AddTagHelper

-0.61

+:+

-0.61

TagMode

-0.61

 ujednoznacz

-0.55

enumi

-0.54

✭✭

-0.53

AutoScale

-0.52

kmal

-0.52

quitous

-0.50

POSITIVE LOGITS

 middle

2.03

middle

1.73

 upper

1.58

 central

1.55

 center

1.40

top

1.37

 MIDDLE

1.35

upper

1.32

 centre

1.30

central

1.28

Activations Density 1.054%

patterns and relationships in logical arrangements involving placements or classifications

like "middle", "top", "central", and "right"

upper and middle positions

No Comments

No Known Activations